sensitivity.result¶

The result of a sensitivity analysis.

SensitivityResult wraps an xarray.Dataset with one variable per observable and index, <observable>.<index> (e.g. auc.ST, auc.ST_conf, [S2].normalized), over (parameter, *other dimensions, [time]): the parameters of the design, every label of the other dimensions of the scan (e.g. doses, conditions) and, for a timecourse on a grid, every time point. attrs carry the units of every variable, the method, its options and the provenance of the scan.

SensitivityResult ¶

SensitivityResult(ds)

The indices of a sensitivity analysis, see the module.

Attributes:

Name	Type	Description
`ds`		the dataset.

Wrap a dataset of indices.

Parameters:

Name	Type	Description	Default
`ds`	`Dataset`	the dataset.	required

Raises:

Type	Description
`ValueError`	if the dataset has no dimension `parameter`.

units `property` ¶

units

Get the unit of every variable, of parameter and of the kept coordinates.

method `property` ¶

method

Get the method of the analysis, e.g. sobol.

parameters `property` ¶

parameters

Get the parameters of the indices.

observables `property` ¶

observables

Get the observables, in the order of their first variable.

index ¶

index(name, observables=None)

Stack an index of the scalar observables into (parameter, observable, ...).

A timecourse is stacked at one time point, e.g. sel(time=10).index(...), which is the state after a change at the time 10, see sel.

Parameters:

Name	Type	Description	Default
`name`	`str`	the index, e.g. `ST`.	required
`observables`	`Sequence[str] \| None`	the observables, every scalar observable which has the index by default.	`None`

Returns:

Type	Description
`DataArray`	The index over `(parameter, observable, *other dimensions)`.

Raises:

Type	Description
`KeyError`	if no scalar observable has the index.
`ValueError`	if `observables` is empty, or a named observable has no such index or is a timecourse.

sel ¶

sel(**indexers)

Select labels, see xarray.Dataset.sel.

A single label of parameter keeps the dimension, which a result has: sel(parameter="k1") is sel(parameter=["k1"]). A single time of a change after the start, which a timecourse has twice, selects the state after the change, see at_one_time; isel selects either.

isel ¶

isel(**indexers)

Select positions, see xarray.Dataset.isel.

A single position of parameter keeps the dimension, as in sel.

to_dataframe ¶

to_dataframe(name)

Get a variable as a table, a row per parameter.

Parameters:

Name	Type	Description	Default
`name`	`str`	the variable, e.g. `auc.ST`.	required

Returns:

Type	Description
`DataFrame`	The table; the other dimensions are its columns.

classify ¶

classify(name)

Classify every value of an index, see sensitivity_classification.

The thresholds are the ones of the IPCS for normalized local sensitivities, which need an index without a unit (DIMENSIONLESS_INDICES), e.g. normalized, S1 or ST.

Parameters:

Name	Type	Description	Default
`name`	`str`	the variable, e.g. `auc.normalized`.	required

Returns:

Type	Description
`DataArray`	The classes as strings (`high`, `medium`, `low`, `negligible`), `""`
`DataArray`	for `NaN`.

Raises:

Type	Description
`ValueError`	if the index has a unit, e.g. `raw` or `mu_star`.

to_netcdf ¶

to_netcdf(path)

Write the result as netCDF, the attributes as JSON.

Parameters:

Name	Type	Description	Default
`path`	`str \| Path`	the file.	required

from_netcdf `classmethod` ¶

from_netcdf(path)

Read a result written by to_netcdf.

Parameters:

Name	Type	Description	Default
`path`	`str \| Path`	the file.	required

Returns:

Type	Description
`SensitivityResult`	The result.

at_one_time ¶

at_one_time(data, indexers)

Select a single time which the data has twice as its state after the change.

The time of a change after the start appears twice in a timecourse, the states before and after the change; a single time selects the state after it, as a time of a grid does (sbmlsim.result.timecourse.grid_weights).

Parameters:

Name	Type	Description	Default
`data`	`T`	the data.	required
`indexers`	`dict[str, Any]`	the labels to select, by dimension.	required

Returns:

Type	Description
`T`	The data at the time, if it has it twice, and the other indexers; else
`dict[str, Any]`	the data and the indexers.

sensitivity.result¶